A statistical method for quantifying the different fouling effects of three combined water sources on an ultrafiltration membrane Original Research Article

A simple statistical model has been developed to explain the fouling of an ultrafiltration (UF) membrane treating three mixed feedwater matrices. A tubular UF membrane operating under dead-end conditions, based at the Millennium Dome Water Recycling Plant, was used for this study. The feedwaters comprised surface water, greywater and groundwater, all having distinct quality differences. The analysis utilises the least-squares fit regression technique. Fouling, as manifested in the diurnal change in trans-membrane pressure (TMP), is related numerically to the volumes of the feedwaters, the mean TMP and the cross product of mean TMP with water volume. It was found that the impact of the recovered greywater on TMP was not statistically significant only due to the low daily volumes of this matrix generated, compared with those of the other sources. The model that best described the data was chosen on the basis of highest R2 and lowest residual mean square, and demonstrated the surface water to be more fouling than the ground water by a factor, which is a function of the mean TMP.